Offset Of Automotive Safety Glass, Vice Like Electronic Detection System

On testing the value of the deviation of subsidiary image of safety glazing materials used on road vehicles, the conventional method is an optical-mechanic one. This method has the disadvantages of low measuring precision, requiring environment of darkroom, and making eyes tired after a long-time testing.With the development of the computer science and micro-electronic technology, charge coupled device (CCD) and image manipulation technology have been widely used in conventional optics testing. Compared with conventional methods, CCD is more sensitive to light and has more precise ability of distinguishing than man's eyes. At the same time, adopting CCD imaging and software testing can also avoid subjective errors to come to more precise testing results. This thesis is mainly focused on a method of real time data acquisition and processing, with the aids of CCD cameras and computer software, which is used to process the value of the deviation of subsidiary image of safety glazing materials used on road vehicles.First, we design the optical system to test the value of the deviation of subjective image. This system is the optical-collimate one, with safety glazing materials fixed on the sample glass bracket from the life-size angle that can move in horizontal and vertical ways. The collimate lens can image the coordinate system with a bright point in the center into the infinitude end. When the sample glass forming the subsidiary image is placed from the life-size angle between CCD and collimate lens, a subsidiary and weaker bright point emerges somewhere in a distance away from the center of coordinate system. Then CCD and Data Acquisition System are used to transform the optical images into computer recognizable data patterns.Software processing of the image in the later stage mainly includes detaching the background, threshold, getting rid of noises, testing target with Hough transform, obtaining the value of the deviation of subsidiary image with the use of model matching, the real time measuring of the subsidiary image and other digital and image processing. In this system, the method we adopt to work out the magnification ratio is to detach the circle of certain radius in the coordinate system of the divisional board and then to obtain the magnification ratio of the coordinate system through the circle radius of the screen. This testing of circle radius adopts the method of Hough transform. As the demarcation error which is brought about in theprocess of obtaining the magnification is the main error made in the testing system, before Hough transform we should make an rough detachment of the image formed by divisional boards, then threshold after getting rid of noises, then detach the image formed by coordinate system and finally skeleton it. This method can greatly reduce the numbers of pixels participating in Hough transform, promote its efficiency, improve the measuring precision and avoid mistesting the circle we do not need. When testing the value of the deviation of subsidiary image, we take in the method of threshold and model matching, which improves the arithmetic efficiency and meets the demands of real-time testing.The experiment results show, comparing with the conventional ones, this method has the advantages of high measuring precision, real time display of result and measurability in the condition of interior illumination.